Epilepsy is a debilitating disease which causes sudden, unpredictable episodes of loss of consciousness or abnormal, automatic behavior. Epileptic individuals are at high risk for injuries by losing consciousness and falling, or during automatisms. It is often impossible for epileptic individuals to be left alone because of the possibility of unattended injury or life-threatening prolonged loss of consciousness and repetitive seizures (status epilepticus). We propose to develop a small, wearable non-invasive device which will trigger an alarm through existing medical alert infrastructure when an individual suffers a seizure. The seizure detector will not require electrical connections to the patient and will recognize events by utilizing multiple miniature accelerometers to sense patients' movements and by real time microcomputer analysis of transducer signal outputs. The unit will also be programmable so that an individual's specific seizures can be detected with enhanced accuracy. We propose to record and analyze motion data from patients of both sexes and various ethnic and age groups. State of the art signal processing techniques will be utilized to develop algorithms which reliably detect seizures while minimizing "false alarms," which will be handled by delayed alert output and manual override. PROPOSED COMMERCIAL APPLICATIONS: The incidence of epilepsy is estimated at more than 1% of U.S. and world populations, with at least 15% of patients suffering intractable seizures. Accordingly, candidates for the seizure detector within the U.S. are estimated to number 370,000 and within developed nations at least 1,000,000.